On 2016-08-16, at 11:12, Jesse 1 Robinson wrote:

> I think that the problem with writing to *any* data set on whatever medium is 
> that it cannot be read while it's open for output. It seems that the log file 
> must be written to sysout (for concurrent examination) and then captured to a 
> data set after it's closed. 
>  
That may depend on whether you're willing to write to the data set
with DISP=SHR.  RACF can protect you a lot.  Allocate DISP=NEW in
an earlier job step; lately you can (Admins willing) downgrade the
ENQ to SHR for the later job step.

Decades ago I used ISPF Browse to view data sets being written.
Sporadic I/O errors; buffering latency; need to END and re-enter
to refresh.  Processors are so fast now (and SDSF exists) that
I'n no longer tempted to do that.

If the client has or is willing to pay for SDSF a following step
using the Rexx (or other) API can capture the SYSOUT.  The SDSF
Ref/Guide has a useful example.  Ironically (absurdly) as Ed J.
has noted, the hardest part may be for a job to discern its own
Job ID.

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to